Drawing a border around a cluster in pair-wise correlation plot

[ghost]

Is there a way to draw a border around a cluster of my choice in a heatmap like this using ComplexHeatmap?:

I have thus far figured out:

draw heatmap using pair-wise pearson correlation plot:

hm_corr <- Heatmap(
  matrix = coeffs,
  name = "hm_corr_test",
  show_column_names = F,
  show_row_names = F,
  width = ncol(coeffs) * unit(0.2, "mm"),
  height = nrow(coeffs) * unit(0.2, "mm"), 
  row_gap = unit(0, "mm"), column_gap = unit(0, "mm")
)

hm_corr = draw(hm_corr)

find the order of the row and column from the heatmap:

rowlist <- row_order(hm_corr)
collist <- column_order(hm_corr)

order the original matrix based on this row and column orders:

coeffs.ordered <- coeffs[rowlist, ]
coeffs.ordered <- coeffs.ordered[, collist]

Make a cluster based on blocks that have >= 0.8 pearson correlation coefficient:

coeffs.ordered.df.c1 <- coeffs.ordered.df[coeffs.ordered.df[, 1] >= 0.8, ]
column_without_number <- names(coeffs.ordered.df.c1)[which(!apply(coeffs.ordered.df.c1 == 1, 2, any))[1]]
number.to.cut <- which(colnames(coeffs.ordered.df.c1) == column_without_number)

Define my own clusters

vector <- c()
cluster1_value <- "cluster 1"
vector <- c(vector, rep(cluster1_value, number.to.cut))
cluster2_value <- "cluster 2"
vector2 <- c(vector, rep(cluster2_value, nrow(coeffs.ordered.df) - number.to.cut))

df <- data.frame(clusters.row = vector2, clusters.col = vector2)
df$clusters.row = factor(df$clusters.row, levels = c("cluster 1", "cluster2"))
df$clusters.col = factor(df$clusters.col, levels = c("cluster 1", "cluster2"))

hm_corr_ordered = Heatmap(
  matrix = as.matrix(coeffs.ordered.df),
  name = "hm_corr_test.ordered",
  show_column_names = F,
  show_row_names = F,
  width = ncol(coeffs) * unit(0.2, "mm"),
  height = nrow(coeffs) * unit(0.2, "mm"), 
  column_split = df,
  row_split = df,
  cluster_column_slices = F,
  cluster_row_slices = F
  
  #row_gap = unit(0, "mm"), column_gap = unit(0, "mm")
)

hm_corr_ordered = draw(hm_corr_ordered)

this gave me

but when I try to use decroate_heatmap_body function, it gives me this error: Viewport 'hm_corr_ordered_heatmap_body_cluster 1_cluster 1' was not found

decorate_heatmap_body("hm_corr_ordered", row_slice = 1, column_slice = 1, {
            grid.rect(gp = gpar(fill = "transparent", lwd = 2))
})

I am assuming I'm not defining the slice correctly. What am I doing wrong? I'd greatly appreciate any advice.

[ghost]

resolved this issue here: #1085